OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: more grist

> > However, tying it to something like the Infoset, and
> > through that, the sort of things you might see in XML Schema, or TREX,
> RELAX, might
> > make sense. That way, the implementation is tied to no specific syntax,
> > while maintaining the features that schemas offer.
> What I find confusing, is that all the newer drafts are clearly reliant
> on the PSVI. So, if this is going to be the 'real' or 'core' Infoset, why
> isn't
> it documented as such?

As Henry Thompson recently told us in London, the important information that
the post schema validation infoset adds to the post parse (or pre schema
validation) infoset is types. Think of schema validation then as simply an
XSLT which when applied to the infoset, adds types and tags indicating
validity. It's really surprisingly simple.

What this also says, and what should (IMHO) be strongly emphasized, is that
any other application can apply its own types via whatever mechanism it
chooses. Such an application might be embodied as:

1) a SAX filter
2) something that operates on a DOM
3) something that operates on any in memory structure that represents the
document such as an RDF model etc.

For example: RDF and RDF Schema has a perfectly good notion of types and I
my personal view an RDF parser could equally supply types to an infoset. At
some might I may write this up in an article, but for the moment take a look
http://www.openhealth.org/XSet which defines a full XML grove syntax -- from
which an Infoset serialization syntax can be whittled away -- and combine
this with
http://www.openhealth.org/RDF/rdfExtractify.xsl which is an XSLT based RDF
parser which implements a simplfied RDF syntax, and parses arbitrary XML
documents into edge labeled graphs -- such an edge labeled graph represents
a subset of the full XML grove for any particular document.

Simply consider this DLG as a grove or infoset, and see that transforms on
this graph are transforms on the Infoset. XML Schema's typing function
becomes alot less ominous when viewed in this fashion (and notice that RDF
already defines this function!!!)